基于机器学习和图像识别的电力作业现场安全监督方法

Safety Supervision Method for Power Operation Site Based on Machine Learning and Image Recognition

针对电力作业现场人员误入危险区域的安全问题,开展人员闯入检测的研究,首先利用梯度方向直方图(histogram of oriented gradient,HOG)和支持向量机(support vector machine,SVM)进行完全帧的人员检测,然后利用基于OpenCV的图像处理技术判断人员是否闯入警戒区域。通过视频监控设备采集作业现场图像,采用上述方法实时识别现场人员及其危险行为,并发出告警信号。实验结果表明,检测结果准确率达到92%,实现了电力作业现场安全监督自动化,显著提升了作业现场安全水平。

Aiming at the safety supervision of power operation site,a research is carried out on personnel intrusion detection.It is proposed to use the histogram of oriented gradient(HOG)and support vector machine(SVM)for full frame personnel detection,and then the OpenCV-based image processing technology is used to determine any personnel intrusion into the alert area.Through the images of the work site captured by video monitoring equipment,the above-said methods are used to identify the on-site personnel and their dangerous behaviors in real time,and to issue an alarm signal.The experimental results show that the accuracy of the detecting results reaches 92%,and the automatic safety supervision of power operation site is realized,which greatly improves the safety level of the power operation site.